High real-time decision method of wearable weak signal

Abstract

The invention relates to a high real-time decision method of wearable weak signal, comprising the following steps of: (1) using a scroll chaotic system as a calculation engine; (2) determining controlparameters of the scroll chaotic system; (3) fusing the characteristic data with the control parameters of the scroll chaotic system and taking the fusion result as the initial value; (4) starting asimulated chaotic hardware circuit, and performing hardware evolution calculation on the initial input value; (5) terminating the hardware evolution calculation with the cooperation of a sequential circuit; (6) starting a hardware detection circuit to detect the output of the simulated chaotic trajectory, and classifying the data according to the comparison result with the voltage at the equilibrium point of the scroll chaotic system; and (7) feeding back the output result of the classification to the digital signal processor to make decisions. According to the high real-time decision method of wearable weak signal, scroll simulation chaos is utilized as a calculation engine, a traditional neural network operation is replaced by hardware implementation, and the problems of many parameters,long training time, much stored data and the like existing in the neural network and influencing the decision response speed are solved.

Description

technical field [0001] The invention belongs to the field of intelligent wearable device control, and in particular relates to a high real-time decision-making method for wearable weak signals. Background technique [0002] As a member of strategic emerging industries, smart wearable devices have been promoted to the level of a new growth engine for the national economy, but there is currently a dilemma of "high user abandonment rate and insufficient rigidity". The reason is that, on the one hand, wearable devices are oriented to a multi-state, multi-environment, long-term, and highly personalized complex application background, with a huge amount of computing tasks; on the other hand, based on considerations such as wearing comfort and aesthetics, The computing resources of wearable devices are severely scarce and cannot respond to events in real time. These factors make it difficult for current wearable products to touch the pain points of users and become tasteless. [...

AI Technical Summary

Problems solved by technology

[0005] Judging from the currently retrieved literature, there are few literature reports on the use of multi-scroll analog chaotic circuits instead of traditional neural networks to achieve data classification.

In addition, most of the existing chaotic computing-based systems are double-scroll or digital chaotic systems. In comparison, multi-scroll chaotic systems have more complex chaotic dynamics, which is convenient for multiple categories to distinguish

Images